Using highly purified amino acids of high specific radioactivity we have been and continue to measure the frequency with which one amino acid inadvertently falls into a position properly intended for another. In two loci in rabbit hemoglobin this frequency of error is quite close to one mistake in 3000. The work is being extended to other pairs of amino acids, to other proteins and to tissues of varying age.